In order to provide upper-layer applications with continuous message delivery and updated information on membership changes -due to process failures, leaves and joins -, group communication protocols must exchange messages and continuously monitor all group members, which in certain load conditions may incur unacceptable message overhead to the underlying communication system. Therefore, the design of such protocols must deal with the trade-off between performance requirements such as speed (e.g., delivery latency) and cost (e.g., message overhead). However, when the behavior of the computing environment is unknown and can change over time, or when the application requirements can dynamically change, self-configuring is a basic issue that is usually ignored in existing implementations. In this paper we present the design, implementation and evaluation of a novel self-manageable group communication protocol based on the feedback control theory, which is capable of self-configuring its operation parameters at run-time from previously specified requirements such as resource consumption.

Castro and Liskov proposed in 1999 a successful solution for byzantine fault-tolerant replication, named PBFT, which overcame performance drawbacks of earlier byzantine faulttolerant replication protocols. Other proposals extended PBFT with further optimizations, improving PBFT performance in certain conditions. One of the key optimizations of PBFT-based protocols is the use a request batching mechanism. If the target distributed system is dynamic, that is, if its underlying characteristics change dynamically, such as workload, channel QoS, network topology, etc., the configuration of the request batching mechanism must follow the dynamics of the system or it may not yield the desired performance improvement. This paper addresses this challenge by proposing an innovative solution to the dynamic configuration of request batching parameters inspired on feedback control theory. In order to evaluate its efficiency, the proposed solution is simulated in various scenarios and compared with the original version used in the PBFT-family protocols.

No abstract

No abstract